Year 12 | 26 January 2020 | staff@teatronaturale.com TO ENTER | TO REGISTER

Nitrofurans and their metabolites in food

Nitrofurans are not authorised for use in food-producing animals in the European Union (EU), but furazolidone, nitrofurantoin and nitrofurazone may be used in human medicine

Nitrofurans are synthetic broad spectrum antimicrobial agents. The nitrofurans considered in this opinion are furazolidone, furaltadone, nitrofurantoin, nitrofurazone and nifursol. Nitrofurans are not authorised for use in food-producing animals in the European Union (EU), but furazolidone, nitrofurantoin and nitrofurazone may be used in human medicine.

Nitrofurans share a nitrofuran ring which is coupled to a side-chain via an azomethine bond. The side-chains differ for the various drugs, being 3-amino-2-oxazolidinone (AOZ) for furazolidone, 3-amino-5-methylmorpholino-2-oxazolidinone (AMOZ) for furaltadone, 1-aminohydantoin (AHD) for nitrofurantoin, semicarbazide (SEM) for nitrofurazone, and 3,5-dinitrosalicylic acid hydrazide (DNSH) for nifursol. Nitrofurans have short half-lives in animals and therefore they do not occur generally as residues in foods of animal origin. Reactive metabolites are formed that are able to bind covalently to tissue macromolecules, such as proteins and DNA. When animal tissues are consumed as food, the side-chains may be released from the metabolites, namely AOZ, AMOZ, AHD, SEM and DNSH.

The EFSA Scientific Opinion, titled ‘Guidance on methodological principles and scientific methods to be taken into account when establishing Reference Points for Action (RPAs) for non-allowed pharmacologically active substances present in food of animal origin’, identified an approach for establishing RPAs for various categories of non-allowed pharmacologically active substances. However, the opinion also identified certain categories of non-allowed pharmacologically active substances that are considered to be outside the scope of the procedure, including substances that are high potency carcinogens, such as nitrofurans. As nitrofurans are excluded from that opinion, and taking into account that the presence of SEM in food may be from sources other than use of nitrofurazone, the European Commission (EC) requested the European Food Safety Authority (EFSA) for a scientific opinion on the risks to human health related to the presence of nitrofurans and their metabolites in food. The opinion should include (a) an evaluation of the toxicity of nitrofurans and their metabolites for humans, considering all relevant toxicological endpoints and identification of the toxicological relevance of nitrofurans and their metabolites present in food, and (b) an exposure assessment of the EU population to nitrofurans and their metabolites from food, including the consumption patterns of specific (vulnerable) groups of the population. In addition, the opinion should assess the appropriateness of using marker metabolites of nitrofurans for the reference point for action for food of animal origin. The opinion should evaluate whether a reference point for action of 1 µg/kg for nitrofuran metabolites, as defined in legislation, in food of animal origin is adequate to protect public health, and it should assess the appropriateness of applying the reference point for action, considered adequate to protect public health, to other commodities than food of animal origin.

Because the nitrofuran parent compounds can only be detected in animal tissues and products for a short period after treatment of the animals, monitoring of nitrofuran residues in livestock based on the identification of the parent compounds is not appropriate. Metabolites binding covalently to proteins and persisting for several weeks in edible tissues, from which the side-chains AOZ, AMOZ, AHD, SEM and DNSH may be released, serve as excellent marker metabolites for the illicit use of nitrofurans in food-producing animals. Generally, both screening and confirmatory methods for the nitrofuran marker metabolites in foods of animal origin use acid hydrolysis and nitrobenzaldehyde derivatisation of the released marker metabolites. Screening for the resulting nitrophenyl derivatives is generally undertaken by enzyme-linked immunosorbent assays (ELISA) or biosensor methods, providing sufficient analytical sensitivity to meet the current minimum required performance limit (MRPL) of 1 µg/kg. Confirmatory methods are based on liquid chromatography–tandem mass spectrometry (LC-MS/MS) and also adequately meet the MRPL of 1 µg/kg.

The EFSA Panel on Contaminants in the Food Chain (CONTAM Panel) concluded that, since other nitrofuran metabolites that could persist at higher concentrations have not been identified, the marker metabolites AOZ, AMOZ, AHD, SEM and DNSH are appropriate as the RPA for foods of animal origin.

Data on occurrence of nitrofuran metabolites (AOZ, AMOZ, AHD and SEM) in food, reported by Member States from the National Residue Monitoring Plans, have been extracted for the period 2002 to 2013; there were 214 non-compliant targeted samples reported for nitrofurans over that 12 year period. The categories in which nitrofurans were reported in decreasing level of incidence were poultry, bovines, sheep/goats, pigs, farmed game, honey, rabbit, aquaculture, horses and wild game. Data were extracted also from the Rapid Alert System for Food and Feed (RASFF) database for the years 2002 to 2014; there were 808 notification events reported for nitrofuran metabolites (AOZ, AMOZ, AHD and SEM), of which 416 were for crustaceans and products thereof and 150 were for poultry meat and poultry meat products.

The CONTAM Panel concluded that data extracted from the EC database and the RASFF database were too limited to carry out a reliable human dietary exposure assessment. Instead, the CONTAM Panel calculated the hypothetical human dietary exposure for a scenario in which foods of animal origin, excluding milk and dairy products, are considered to contain one nitrofuran marker metabolite at a concentration level equal to the RPA of 1 µg/kg. This scenario, representing a worst-case situation for the occurrence of nitrofuran marker metabolites due to illicit nitrofuran use, is a highly unlikely situation. The mean chronic dietary exposure across dietary surveys for this scenario would range from 1.9 to 4.3 ng/kg b.w. per day for adults and would be the highest for toddlers, at 3.3 to 8.0 ng/kg b.w. per day.

Besides arising from nitrofurazone use, SEM may occur in food from other sources, including use of the food additive carrageenan. The CONTAM Panel considered scenarios covering the different sources. In one exposure scenario, foods of animal origin (including only those milk and dairy products for which carrageenan is authorised as an additive) and foods of non-animal origin for which carrageenan is authorised as an additive, were included. These foods are considered to be contaminated with SEM at a concentration level equal to the RPA of 1 µg/kg; this scenario covers all potential dietary exposure. The mean chronic dietary exposure to SEM across dietary surveys for this scenario would range from 6.4 to 16 ng/kg b.w. per day for adults and would be the highest for toddlers, at 17 to 55 ng/kg b.w. per day.

Reduction of the nitro group seems to be the most important metabolic pathway for nitrofurans, potentially leading to reactive intermediates that are capable of binding to proteins and to DNA. Nitroreduction and subsequent redox-cycling results in the generation of reactive species (including oxygen species) that might be responsible for some of the adverse effects.

Based on studies with radiolabelled nitrofurans, high levels (mg/kg range) of metabolites are present in tissues shortly after the last treatment. A proportion of the metabolites cannot be extracted from the tissues with organic solvents and are assumed to be protein-bound. Levels of these residues decrease gradually but are still detectable after 45 days in muscle, kidney and liver of treated pigs and probably for much longer. The decrease of residues in liver and kidney is faster than in muscle tissue.

Feeding of rats with protein-bound residues of radiolabelled furazolidone showed that some of the radiolabel was excreted in urine and so must have been absorbed in the gastrointestinal tract. The radiolabel was also detected in tissues of rats and was partly non-extractable. AOZ could be released by acid treatment from these non-extractable residues in rat tissues. Free AOZ was detected in blood of rats fed with meat containing only protein-bound residues of furazolidone, showing that AOZ can also be released from these residues, probably due to acid hydrolysis in the stomach.

Acute toxicity studies in laboratory animals showed that for furazolidone, nitrofurantoin and nitrofurazone the lung is an important target for toxicity, leading to decreased respiratory function and death. Signs of neurotoxicity such as hyperirritability, tremors and convulsions were also found.

In repeated dose toxicity studies, AOZ caused hepatotoxicity, decreased body weight gain and anaemia at the lowest tested dose of 0.9 mg/kg b.w. per day in rats and at 1 mg/kg b.w. per day in dogs. Nitrofurantoin caused toxic effects in liver, kidney and testes, and caused necrosis of the ovarian follicles, decreased weight gain and neurotoxicity, with a no observed adverse effect level (NOAEL) of about 120 mg/kg b.w. per day in rats and mice. Nitrofurazone caused similar effects as nitrofurantoin, with the exception of necrosis of the ovarian follicles, and the NOAEL for effects on the testes in rats was 13.5 mg/kg b.w per day. SEM caused severe deformation of limbs and osteochondral lesions at the lowest tested dose of 23 mg/kg b.w. per day in rats. Nifursol caused slight changes in red blood cell parameters and a NOAEL of about 14 mg/kg b.w. per day was identified.

In studies on spermatogenesis, furazolidone, furaltadone, nitrofurantoin and nitrofurazone caused toxic effects on the testes in rats and mice but no NOAEL could be identified. Effects were observed at the lowest dose tested of 10 mg/kg b.w. per day for nitrofurantoin.

In studies on embryotoxicity and teratogenicity, furazolidone in mice was embryotoxic at the lowest dose tested of 200 mg/kg b.w. per day and caused decreased body weight and viability of pups after birth, but no malformations were found. Nitrofurantoin was embryotoxic in mice and rats and caused decreased body weight and viability of pups after birth. A NOAEL of 10 mg/kg b.w. per day was identified for embryotoxicity in rats. Malformations were not found in offspring of rats and rabbits, with a NOAEL of 30 mg/kg b.w. per day for teratogenicity. Nitrofurazone was not teratogenic in mice and rabbits at doses that were not maternotoxic. For fetotoxicity/maternotoxicity an overall NOAEL of 14 mg/kg b.w. per day was identified. For SEM, in a study looking at the incidence of cleft palate and resorptions only, an effect was found when rats were treated orally with SEM at 25 mg/kg b.w. per day or higher, but not when treated at 10 mg/kg b.w. per day.

In multigeneration studies, nitrofurazone showed reproductive toxicity in mice for two generations at doses of 14 to 102 mg/kg b.w. per day. Nifursol did not have any effects on reproduction in rats treated for three generations at doses of 54 mg/kg b.w. per day or lower.

In studies on neurotoxicity, nitrofurantoin caused peripheral nerve damage in rats treated orally at the lowest dose tested of 20 mg/kg b.w. per day. SEM caused neurobehavioural effects in juvenile rats when treated orally at the lowest dose tested of 40 mg/kg b.w. per day for 10 days.

In genotoxicity studies, furazolidone and its marker metabolite AOZ were found to be genotoxic in vitro and possibly also in vivo. Since AOZ can be released from bound residues of furazolidone metabolites, these bound residues should be considered as genotoxic. Furaltadone was found to be a bacterial and mammalian cell mutagen in vitro. The marker metabolite AMOZ is not genotoxic in vitro. In vitro, nitrofurantoin induces mutations, chromosomal aberrations and DNA damage and, in vivo, nitrofurantoin has been shown to induce DNA damage in multiple organs, micronuclei formation in mice and gene mutations in a transgenic mouse mutation assay. For AHD, the only in vivo mutagenicity study which is available shows a negative result. Nitrofurazone and its marker metabolite SEM are genotoxic in vitro. In vivo tests gave negative results with nitrofurazone, whereas no conclusion can be drawn on the in vivo genotoxicity of SEM. Nifursol is genotoxic in vitro, whereas in vivo it induced neither chromosomal aberrations nor mutations.

In chronic toxicity and carcinogenicity studies, furazolidone induced malignant mammary tumours in rats, bronchial adenocarcinomas in male and female mice and neural astrocytomas in male rats. The CONTAM Panel concluded that furazolidone is carcinogenic in mice and rats. No information on the carcinogenicity of AOZ, the marker metabolite of furazolidone, was identified, but it is presumed that AOZ may play a role in tumour formation. Furaltadone induced malignant mammary tumours in female rats. The CONTAM Panel concluded that furaltadone is carcinogenic in rats. There is no information on the chronic toxicity or the carcinogenicity of AMOZ. Nitrofurantoin induced an increase mainly in benign tumours in mice and rats, but in male rats a few malignant tumours were found. Based on these observations, the CONTAM Panel concluded that there is limited evidence that nitrofurantoin is carcinogenic in rats. No information on the chronic toxicity or the carcinogenicity of AHD was identified. Nitrofurazone increased the incidence of mainly benign tumours in mice and rats following oral administration. In male rats a non-dose related increase in carcinomas of the preputial gland was observed. The CONTAM Panel concluded that there is no evidence for the carcinogenicity of nitrofurazone in mice, and that evidence for its carcinogenicity in rats is equivocal. Non-neoplastic effects of nitrofurazone were observed in a chronic toxicity study at the lowest dose tested of 14 mg/kg b.w. per day in mice (ovarian atrophy in females and reduced survival in males) and the lowest dose tested of about 11 mg/kg b.w. per day in rats (testes degeneration). SEM increased the incidence of malignant lung tumours, particularly in female mice. In rats, no increase in tumour incidence was found. The CONTAM Panel concluded that there is limited evidence that SEM is carcinogenic in mice, but not in rats. Based on effects on bones observed in a chronic toxicity study in male rats, a NOAEL of 0.6 mg/kg per day was derived for non-neoplastic effects of SEM. For nifursol the available chronic toxicity studies in rats and dogs did not show clear indication for carcinogenicity. The toxicological information was too limited to derive a NOAEL for non-neoplastic effects of nifursol. No information on the chronic toxicity or the carcinogenicity of DNSH was identified.

In relation to the mode of action, reduction of the nitro-group seems to be the key metabolic pathway leading to reactive intermediates, including reactive oxygen species. Reactive metabolites are capable of binding to proteins and to DNA, being thereby responsible for most of the adverse effects resulting from exposure to nitrofurans. Only for AOZ information was identified regarding the mode of action of the nitrofuran marker metabolites. AOZ plays a role in the inhibition of monoamine-oxidase in animals treated with furazolidone. This may result in an increased susceptibility to neurotoxic effects of certain biogenic amines such as tyramine. Protein binding of reactive nitrofuran metabolites may play a role in the irreversible inhibition of the pyruvate dehydrogenase complex, another potential mechanism underlying neurotoxic effects of nitrofurans, such as polyneuritis.

In human studies, oral administration of furazolidone and nitrofurantoin may lead to a range of adverse reactions, particularly nausea, vomiting and abdominal pain. Both drugs have also been associated with haemolytic anaemia observed in patients deficient in glucose-6-phosphate dehydrogenase. The topical use of nitrofurazone may lead to allergic reactions. Epidemiological studies are reported only for patients treated with nitrofurantoin, and associations were found for cancers of the nervous system in adults, for drug-induced liver injury, and for increased risk of pulmonary adverse events in patients with renal impairment.

Because most of the nitrofurans and their marker metabolites are genotoxic and/or carcinogenic, derivation of health-based guidance values (HBGVs) is not appropriate.

In the case of furazolidone, a lower 95 % confidence limit for a benchmark response of 10 % extra risk (BMDL10) value for bronchial adenocarcinomas in mice of 3.5 mg/kg b.w. per day (1.6 mg/kg b.w. per day, expressed as AOZ) was selected as a reference point for carcinogenic effects. Non-neoplastic effects of furazolidone and AOZ were found on red blood cell parameters and enzymes in blood. The lowest BMDL was estimated for the effect of AOZ on alkaline phosphatase (ALP) (BMDL05 of 0.02 mg/kg b.w. per day). The CONTAM Panel concluded that this value can be used as reference point for the risk characterisation for non-neoplastic effects.

For furaltadone, the CONTAM Panel concluded that the available data do not provide a suitable basis for deriving a reference point. For AMOZ there is no information on carcinogenicity, and the limited available data indicate that it is non-genotoxic in vitro. Therefore, the CONTAM Panel concluded that the risk for carcinogenicity cannot be assessed. There is no information on non-neoplastic effects of furaltadone or AMOZ that could be used for the derivation of a reference point.

In the case of nitrofurantoin, a BMDL10 value for osteosarcomas in male rats of 61 mg/kg b.w. per day (29.5 mg/kg b.w. per day, expressed as AHD) was selected as a reference point for carcinogenic effects. For non-neoplastic effects, the most sensitive endpoint for nitrofurantoin is impaired spermatogenesis, but the available data did not allow for a BMD analysis or the derivation of a NOAEL. Effects were observed at the lowest dose tested of 10 mg/kg b.w. per day (4.8 mg/kg b.w. per day, expressed as AHD) and this was selected as a reference point for non-neoplastic effects. The CONTAM Panel noted that the effects at this dose are substantial.

For nitrofurazone, no conclusion could be drawn on its possible carcinogenicity and in the case of SEM, the available information was not suitable to derive a reference point for carcinogenic effects. Non-neoplastic effects of nitrofurazone were found on the testes and the epididymis in rats, while for SEM effects on bone development were observed. The lowest BMDL was estimated for the effect of SEM on bone development (BMDL10 of 1.0 mg/kg b.w.). The CONTAM Panel concluded that this value can be used as reference point for the risk characterisation for non-neoplastic effects.

While nifursol is genotoxic in vitro, there is no clear indication that it is carcinogenic and for DNSH there is no information on mutagenicity/genotoxicity or carcinogenicity. For non-neoplastic effects, a BMDL05 value for the effect of nifursol on liver weight of 11 mg/kg b.w. per day (7.3 mg/kg b.w. per day, expressed as DNSH) was selected as reference point.

Since different critical effects are observed for the different marker metabolites, the CONTAM Panel characterised the risk for each marker metabolite separately. For the actual exposure to nitrofuran marker metabolites, no reliable human dietary exposure assessment could be carried out and, therefore, the CONTAM Panel could not characterise the risk.

To evaluate whether the RPA for nitrofuran metabolites in food of animal origin is adequate to protect public health, the CONTAM Panel considered the scenario in which foods of animal origin, excluding milk and dairy products, are considered to contain one nitrofuran marker metabolite at a concentration level equal to the RPA of 1 µg/kg.

For AOZ, median chronic dietary exposure across dietary surveys for the average consumer would result in a margin of exposure (MOE) for carcinogenicity of about 2.9 × 105 for toddlers and 6.2 × 105 for adults and an MOE for non-neoplastic effects of about 3.6 × 103 for toddlers and 7.7 × 103 for adults. The CONTAM Panel considered that for AOZ these MOEs for carcinogenicity and non-neoplastic effects are sufficiently large and do not indicate a health concern.

For AMOZ, the CONTAM Panel could not conclude on the carcinogenicity. Given that there are no clear indications that furaltadone is more potent than furazolidone with respect to the induction of mammary adenocarcinomas, the CONTAM Panel concluded that the cancer risk from AMOZ, if any, would not be greater than that from AOZ and hence does not indicate a health concern. The CONTAM Panel could not identify a reference point for non-neoplastic effects for AMOZ.

For AHD, median chronic dietary exposure across dietary surveys for the average consumer would result in an MOE for carcinogenicity of about 5.4 × 106 for toddlers and 1.1 × 107 for adults and an MOE for non-neoplastic effects of about 8.7 × 105 for toddlers and 1.8 × 106 for adults. The CONTAM Panel considered that for AHD these MOEs for carcinogenicity and non-neoplastic effects are sufficiently large and do not indicate a health concern.

For SEM the cancer risk could not be assessed. For non-neoplastic effects, median chronic dietary exposure across dietary surveys for the average consumer would result in an MOE of about 1.8 × 105 for toddlers and 3.8 × 105 for adults. The CONTAM Panel considered that for SEM these MOEs for non-neoplastic effects are sufficiently large and do not indicate a health concern.

For DNSH, median chronic dietary exposure across dietary surveys for the average consumer would result in an MOE for non-neoplastic effects of about 1.3 × 106 for toddlers and 2.8 × 106 for adults. The CONTAM Panel considered that for DNSH these MOEs for non-neoplastic effects are sufficiently large and do not indicate a health concern.

To assess the appropriateness of applying the RPA that is considered adequate to protect public health to other commodities than food of animal origin, the CONTAM Panel considered the scenario in which foods of animal origin, including only those milk and dairy products for which carrageenan is authorised as an additive, and foods of non-animal origin for which carrageenan is authorised as an additive, are considered to be contaminated with SEM at a concentration level equal to the RPA of 1 µg/kg.

AOZ, AMOZ, AHD or DNSH have not been reported to occur in foods of non-animal origin. Only SEM is reported to occur in food of non-animal origin due to its potential presence in the food additive carrageenan, which is used in a large variety of foods. The food additive carrageenan may also be used in foods of animal origin. For SEM, the cancer risk could not be assessed. For non-neoplastic effects, median chronic dietary exposure across dietary surveys for the average consumer would result in an MOE of about 3.4 × 104 for toddlers and 1.0 × 105 for adults. The CONTAM Panel considered that for SEM these MOEs for non-neoplastic effects are sufficiently large and do not indicate a health concern.

The CONTAM Panel recommends that there is need for a carcinogenicity study on SEM according to the current guidelines and that there is need for information on the mechanisms underlying the genotoxic and carcinogenic effects of SEM.
Keywords

by S. C.
29 june 2015, Food & Fun > Health

MOST POPULAR ARTICLES